Case Study

AI Contract Management Platform

Industry:Financial services

Technologies:ReactJS, Claude, GPT, AWS

Our client came to us looking for a better way to manage their contracts. The executives, lawyers and finance teams didn’t have an easy way to manage contract renewal dates leading to a lot of unnecessary manual work. They needed a system to quickly summarize their contracts, so they could set reminders on key contract dates, quickly understand recurring and one-time charges and more.

Our client in the cybersecurity industry came to us needing an MVP to manage their contracts and make sure contract renewals aren’t missed. Unfortunately when you have multiple contracts it’s hard for legal and finance teams to keep track and important dates get missed.

Our team at Productera were able to take our client’s MVP product vision and turn that into a working product in 1.5 months. Through AI innovation we designed and implemented a solution with over 80% accuracy. Our client now has a contract dashboard to stay on top of their contracts, set reminders for contract renewals and understand key billing frequencies. Ensuring nothing gets missed and their legal & finance teams stay in sync as the company grows.

Valeria Kruchinina

Project Manager, Productera

Challenges

The client needed an MVP solution for their contract management issues and needed it launched within 2 months.Managing contracts became too much of a problem for our client’s company. Working with multiple contract renewal dates, monthly recurring charges, non-recurring charges, was driving their legal & finance team to the brink.

They needed a dashboard to view all of their contracts in one place, be able to set reminders for key renewal/start/end dates and understand billing amounts/frequencies.

Solutions

AI LLM Models: Claude 3.5, GPT4.0, Llama 2.0
Infrastructure: AWS, Terraform, Bitbucket
Front-End Frameworks: ReactJS, Claude, GPT, AWS

This was a perfect opportunity to use AI LLM models like Claude and GPT to help us bring the key contract information into the platform.

We assigned a team of 1 product manager, 2 full-stack developers, 2 data scientists, 1 quality assurance specialist and 1 designer to help our client achieve their product vision. With the tight timeline of 2 months to launch, our product manager brought design work into our Jira product backlog, had weekly sprint planning with our client and used daily standups to keep the engineering team on track towards our product goal.

We created an algorithm to parse contract documents correctly, bringing in the key fields needed for the system including: contract start/end/renewal dates, monthly recurring and non-recurring charges, supplier, location, terms and more.For AI, we tested Llama 2.0 at first, but needed better accuracy as we had a goal of 80% as our key KPI. So we switched to using 2 AI LLM models Claude 3.5 and GPT 4.0 and hit over 80% accuracy with both models.

We also consulted with our CTO to Terraform the DevOps infrastructure inside AWS. We also were using BitBucket to manage our testing and production environments.

Results

We completed the project in 1 month, hitting our sprints every week for a total of 4 releases. Through testing 38 contracts we had an over 80% accuracy rate which was higher than our key KPI goal of 80%.

Explore Our Other Case Study

AI-Powered Compliance Platform

Replacing compliance officers with AI call monitoring Our client in the financial services industry saw an opportunity for a startup with an AI-powered compliance platform. We helped our client build the product from the ground up. Starting with design workshops and managing full stack engineering sprints to build out the compliance platform for content creation, compliance management and research analyst workflows.

View next